yep, saw this video a few months ago. Wasn't a very good test.
1. As newt points out. The physics test is horrible. However this is a result of javascript problem. This would be fine if the game testing was about physics test. But it's not. It's about rendering performance.
2. Rendering performance. C2 would still end up at the lower due to lack of effective browser acceleration on the test machine.
Unfortunately we maybe able to export to IOS and Android C2 does not push a fully encapsulated game on the those target OS. So odds are performance will only at best be comparable to a native app. If we are using CJS.
Though it still points out an important factor. C2 really isn't the best solution for IOS/Android. It has acceptable results when using CJS.
If these tests were on desktop machine the performance would be comparable. However they seemed to designed to single out a particular engine to be on top.